Project success probability

Can you estimate  - even roughly - the probability of success for the ongoing project?

Setting clear project objectives and ensuring smooth communication are generally considered as necessary ingredients for project success. However, this time focus is on estimating the probability of achieving a certain project budget and end date.

Some kind of simulation - such as previously described Monte Carlo method - is required, see also a step-by-step description of the method.

Chosen probability distributions and related parameters for the simulation have to be reliable. Task history can be used if similar projects have been performed previously. When applying Monte Carlo method simulation can be based either on the actual task history or on a fitted probability distribution with the required parameters.

In the first case the actual task history is sampled to get a task cost and duration for each simulation round. These are then used to get total project cost and duration. Second case - fitting the probability distribution and parameters - has been described in a separate blog post.

MonteCarloProject with MonteCrystal plan supports both methods. They give very similar results, if task history resembles closely a certain probability distribution. Simulation can give e.g. following results:

In the case of really unique projects historical data cannot be used. You can then use a so called Delphi method, where a group of experts give their task cost and duration estimates anonymously. Results from the first questioning round are returned to the experts so that estimates can be adjusted. Final cost and duration estimates can then be used in simulation.